See corresponding editorial on page 1121.

Serum vitamin C and the prevalence of vitamin C deficiency in the

United States: 20032004 National Health and Nutrition Examination
Survey (NHANES)1,2
Rosemary L Schleicher, Margaret D Carroll, Earl S Ford, and David A Lacher

INTRODUCTION

Vitamin C (ascorbic acid) is an indispensable cofactor in the

hydroxylation of proline and lysine, and it is essential to collagen
synthesis and connective tissue integrity. Vitamin C functions as
a reducing agent in hydroxylation reactions catalyzed by dopamine
b-monooxygenase and peptidyl glycine a-amidating monooxygenase (1). It plays an important role in increasing the content
of endothelial cell tetrahydrobiopterin and thereby increasing the
activity of nitric oxide synthase (2). It is involved in the biosynthesis of carnitine, histamine, and several adrenal steroids; it
promotes iron absorption and mobilization; and it functions in
tyrosine, folate, and xenobiotic metabolism. When intake of vi-

1252

tamin C is below a critical amount (10 mg/d) for prolonged periods, failure of wounds to heal, petechial hemorrhages, follicular
hyperkeratosis, bleeding gums, and related abnormalities ensue in
a condition known as scurvy (3). Manifest scurvy has rarely been
reported in the United States during the past 30 y (4). Latent
scurvy characterized by fatigue, irritability, vague, dull aching
pains, and weight loss (5) may be underreported because it is not
recognized as such.
Although epidemiologic evidence suggests that vitamin C
rich foods play a protective role against development of cancers
of the mouth, larynx, esophagus, and stomach (68), intervention
studies that used supplements have not shown protective effects
(911). Similarly, epidemiologic studies suggest a lower risk of
coronary heart disease associated with higher intakes of fruit,
vegetables, and whole grains (1214); however, prospective studies
relating cardiovascular disease with the intake of vitamin C or
serum concentrations have provided mixed results (15, 16).
Studies involving food or supplements or both have shown
mixed results for the effects of vitamin C on oxidative damage to
the eye, hypothesized to be part of the pathogenesis of cataract
and macular degeneration. The Age-Related Eye Disease Study
(AREDS), a large randomized trial, showed no benefit of 500 mg
vitamin C/d (together with vitamin E, b-carotene, and zinc) on
the development or progression of cataracts (17). However, twothirds of the study participants were taking multivitamins containing vitamin C; thus, a treatment effect may have been difficult to discern (18). A 5-y prospective study in Japan that
included .700 cases of newly diagnosed cataract found that
higher vitamin C intake (dietary and supplemental) was associated with reduced incidence of age-related cataracts (19).
Fewer studies have tested the ability of vitamin C to delay or
retard the progression of macular degeneration. In the AREDS trial,
the antioxidant formulation (vitamins C and E and b-carotene)
1

From the Division of Laboratory Sciences, National Center for Environmental Health (RLS), the Division of Health Examination Statistics, National Center for Health Statistics (MDC and DAL), and the Division of
Adult and Community Health, National Center for Chronic Disease Prevention and Health Promotion (ESF), Centers for Disease Control and Prevention, Atlanta, GA.
2
Address correspondence to RL Schleicher, Centers for Disease Control
and Prevention, 4770 Buford Highway, MS F55, Atlanta, GA 30341. E-mail:
rschleicher@cdc.gov.
Received September 24, 2008. Accepted for publication July 6, 2009.
First published online August 12, 2009; doi: 10.3945/ajcn.2008.27016.

Am J Clin Nutr 2009;90:125263. Printed in USA. 2009 American Society for Nutrition

Downloaded from www.ajcn.org by on November 6, 2009

ABSTRACT
Background: Vitamin C (ascorbic acid) may be the most important
water-soluble antioxidant in human plasma. In the third National
Health and Nutrition Examination Survey (NHANES III, 1988
1994), 13% of the US population was vitamin C deficient (serum
concentrations ,11.4 lmol/L).
Objective: The aim was to determine the most current distribution
of serum vitamin C concentrations in the United States and the
prevalence of deficiency in selected subgroups.
Design: Serum concentrations of total vitamin C were measured
in 7277 noninstitutionalized civilians aged 6 y during the crosssectional, nationally representative NHANES 20032004. The prevalence of deficiency was compared with results from NHANES III.
Results: The overall age-adjusted mean from the square-root transformed (SM) concentration was 51.4 lmol/L (95% CI: 48.4, 54.6).
The highest concentrations were found in children and older persons. Within each race-ethnic group, women had higher concentrations than did men (P , 0.05). Mean concentrations of adult
smokers were one-third lower than those of nonsmokers (SM:
35.2 compared with 50.7 lmol/L and 38.6 compared with 58.0
lmol/L in men and women, respectively). The overall prevalence
(6SE) of age-adjusted vitamin C deficiency was 7.1 6 0.9%. Mean
vitamin C concentrations increased (P , 0.05) and the prevalence
of vitamin C deficiency decreased (P , 0.01) with increasing socioeconomic status. Recent vitamin C supplement use or adequate
dietary intake decreased the risk of vitamin C deficiency (P , 0.05).
Conclusions: In NHANES 20032004, vitamin C status improved,
and the prevalence of vitamin C deficiency was significantly lower
than that during NHANES III, but smokers and low-income persons
were among those at increased risk of deficiency.
Am J Clin
Nutr 2009;90:125263.

VITAMIN C DISTRIBUTION IN THE UNITED STATES

1253

Design and data collection

Variables

Laboratory methods

Important correlates of vitamin C concentrations presented in

this study include sex, age, race-ethnicity, smoking status, adiposity, socioeconomic status, vitamin C supplement use derived

SUBJECTS AND METHODS

Subjects

In NHANES 20032004, 7277 persons aged 6 y and 4438

adults aged 20 y and in NHANES III 20636 persons aged 6 y

Downloaded from www.ajcn.org by on November 6, 2009

NHANES 20032004 was a complex, multistage, area probability

sample that was representative of the US noninstitutionalized civilian
population during 20032004. Data collection consisted of 3 phases:
1) a screening visit during which sample persons were identified, 2)
an interview during which a wide battery of health-related questions
were asked, and 3) an examination consisting of direct standardized
physical examinations, including body measurements and blood and
urine collections, performed in a mobile examination center (MEC).
Oversampling was done for elderly individuals, adolescents, pregnant women, Mexican Americans, African Americans, and, beginning in the year 2000, low-income non-Hispanic white persons.
Further information on the NHANES survey is available elsewhere
(21). The NHANES 20032004 sample included 10,791 persons
aged 6 y, 8378 of whom were interviewed (77.6%) and 7982 of
whom were interviewed and examined (74.0%). Approximately 9%
(n = 705) of all examined persons had missing data on serum
vitamin C. The NHANES 20032004 sample included 6916
persons aged 20 y, 5041 of whom were interviewed (72.9%) and
4742 of whom were interviewed and examined (68.6%). Approximately 6% (n = 304) of examined adults aged 20 y had missing data
on serum vitamin C in NHANES 20032004.
NHANES III, also a complex, multistage, area probability
sample representative of the US noninstitutionalized civilian
population, was conducted from 1988 to 1994. The NHANES III
sample included 30,930 persons aged 6 y, 25,733 of
whom were interviewed (83.2%) and 23,070 of whom were
interviewed and examined (74.6%) in the MEC. Of these, 11%
(n = 2434) had missing serum concentrations of vitamin C. The
NHANES III sample included 23,258 persons aged 20 y,
18,825 of whom were interviewed (81.0%) and 16,573 of whom
were interviewed and examined (71.0%) in the MEC. Of these,
9% (n = 1388) had missing serum concentrations of vitamin C.
Details of the NHANES III survey design (22, 23) and vitamin C
results for persons 1274 y of age (20) have been published
previously. For most analyses, adults aged 20 y, excluding
children and adolescents, were considered because younger
persons tended to have better vitamin C status than adults.

and 15,185 adults aged 20 y had total serum concentrations of

ascorbic acid (oxidized and reduced) measured. Within 30 min
of separation from the clot, one part serum was treated with 4
parts of 6% metaphosphoric acid in the MEC. The stabilized
serum samples were stored frozen, then shipped to the Centers
for Disease Control and Prevention in Atlanta where they were
stored at 270C until tested. An isocratic reverse-phase HPLC
method with electrochemical detection of ascorbic acid was
used for NHANES 20032004 (24); a similar method was used
for NHANES III (25). The analytic method used for NHANES
20032004 differed from the one used for NHANES III in 2
important ways. For 20032004, but not for NHANES III, calibrators were treated as samples in the assay, ie, subjected to the
same sample preparation steps. Similarly, an internal standard
was used in all sample preparations to correct results for recovery for NHANES 20032004 but not for NHANES III. For
both survey periods, 3 levels of bench quality controls, and
2 levels of blind quality controls, were incorporated into each
assay. In addition, the laboratory participated in semiannual
exercises sponsored by the National Institute of Standards and
Technology (NIST; Gaithersburg, MD) for quality assurance for
vitamin C analysis (26).
Historical quality assurance data were compared to assess
differences in serum vitamin C assay precision with the use of the
2 assay methods. NHANES III bench quality control (QC) data
for 3 pools, ranging in concentration from 21.0 to 88.6 lmol/L,
had CVs between 5% and 8%. Similarly, during NHANES
20032004, 3 bench QC pools, ranging from 13.6 to 122.7
lmol/L, had CVs from 5% to 9%. Blind QC pools (1 in every
20 samples, unknown to analyst, labeled with unique specimen
identification numbers) used in NHANES III ranged in concentration from 5.1 to 48.9 lmol/L and had CVs from 51% to
74%. In NHANES 20032004, the blind QC pool concentrations
ranged from 13.1 to 92.1 lmol/L and had CVs of 59%. Split
sample data (unknown to the laboratory; n = 151 pairs of split
specimens, labeled with different specimen identification numbers) collected during the 6-y course of NHANES III showed an
average paired difference CV of 35%. Split sample data were not
collected for vitamin C during NHANES 20032004. A comparison of the accuracy of methods was made with the use of
NIST standard reference materials (2 levels of SRM 970), which
were analyzed in 1314 assays per method over a period of 2 y
between survey periods while the new method was being validated. With the use of the NHANES III assay method, values
were 9293% of target values compared with 100102% of
target values with the NHANES 20032004 method (24).
With the use of 308 convenience specimens, the average bias
between assay methods was 2.6% with the NHANES 20032004
method, giving higher values than the NHANES III method (24);
however, the bias was concentration dependent with less bias at
lower concentrations. For comparison with NHANES 2003
2004, NHANES III data were adjusted with the use of a Deming
regression equation, where the adjusted values (y) = 1.0566
(original NHANES III) 2 1.9345 lmol/L (24).

plus zinc delayed progression 5 y after the start of therapy (17);

thus, an AREDS-type supplement is currently recommended for
persons with certain stages of macular degeneration.
The National Health and Nutrition Examination Survey
(NHANES) 20032004 is part of the continuous annual survey
conducted by the National Center for Health Statistics, Centers
for Disease Control and Prevention. It provides important information about the consumption of selected nutrients by providing nutritional biomarker measurements. This report presents
the first nationally representative data for serum vitamin C since
NHANES III (19881994) (20) and compares the prevalence of
vitamin C deficiency in the 2 survey periods.

1254

SCHLEICHER ET AL

from a home interview, and dietary intake. Smoking status and

adiposity were measured; all other correlates were self-reported.

dicated yes or no to a question about vitamin-mineral supplement

use in each survey.

Smoking status

Dietary intake of vitamin C

An assessment of tobacco product exposure was based on the

measurement of serum cotinine, the primary proximate metabolite of nicotine (27, 28). Serum cotinine was measured with the
use of a liquid chromatography atmospheric-pressure chemical
ionization tandem mass spectrometry procedure (28). Previous
studies established that ,2% of self-reported tobacco users have
serum cotinine concentrations ,1015 ng/mL (27); thus, individuals with concentrations .10 ng/mL cutoff were classified
as smokers and all others were considered nonsmokers.
For subjects aged 20 y, serum cotinine data from 4434 and
15,053 subjects were available from NHANES 20032004 and
NHANES III, respectively.

Daily intake of vitamin C less than the Estimated Average

Requirement (EAR) was calculated for adults 20 y based on
current recommendations (31) with the use of a single 24-h
dietary recall. For women, EAR values are 60 mg/d for nonsmokers and 95 mg/d for smokers; for men, EAR values are
75 mg/d for nonsmokers and 110 mg/d for smokers. Pregnant
and lactating women have specifically higher values, namely
70 mg/d for pregnant nonsmokers, 105 mg/d for pregnant
smokers, 100 mg/d for lactating nonsmokers, and 135 mg/d for
lactating smokers. Estimates of dietary intake of vitamin C were
available from 95% and 97% of subjects from NHANES 2003
2004 and NHANES III, respectively.

Body mass index

Socioeconomic status
Income status was defined with the use of the poverty-income
ratio (PIR), which is calculated by dividing family income by
a poverty threshold that is specific for family size. This measure
of income has the advantage of being relatively stable over time,
thus enabling comparisons of PIR groups between NHANES III
and NHANES 20032004. Low, medium, and high incomes were
defined as PIR ,1, 1 to ,3, and 3, respectively. PIR values ,1
are below the official poverty threshold, whereas PIR values of
1 indicate income at or above the poverty level (30). Of the
4438 adults in NHANES 20032004, data from 4195 (94.5%)
contained information on PIR; of the 15,185 adults in NHANES
III, 13,779 (90.7%) had information on PIR.
Vitamin C supplement use
For NHANES 20032004 and NHANES III, information
pertaining to the use of nutritional supplements was obtained
during the home interview for persons aged 20 y. In 2003
2004, the question posed was as follows: Have you used or
taken any vitamins, minerals, or other dietary supplements in the
past 30 days? Include prescription and nonprescription supplements. In NHANES III, during the household interview, respondents were asked Have you taken any vitamins or minerals
during the past month? Please include those that are prescribed
by a doctor and those that are not prescribed. In both surveys
those who answered affirmatively were asked to show the interviewer the contents of all vitamin and mineral supplements.
The interviewer then recorded whether any vitamin Ccontaining
supplements were present. More than 99% of respondents in-

Statistical methods
Because the distribution of serum vitamin C data was highly
skewed, a transformation was needed to approximate a Gaussian
distribution to construct CIs and to test statistical hypotheses (32,
33). The square root transformation optimally improved the data
distribution. Estimates of the mean with the use of the square root
transformation; 5th, 10th, 25th, 50th (or median), 75th, 90th, and
95th percentiles; and percentage (and SE) of deficient persons
(vitamin C , 11.4 lmol/L) are presented for persons aged 6 y
and for persons aged 20 y. Sample weights, which account for
unequal probability of selection and adjust for nonresponse and
noncoverage, were incorporated in estimating means, percentiles, percentages, and their standard errors to obtain unbiased
estimates. Standard errors were estimated with the Taylor Series
Linearization, a design-based approach (32).
CIs were constructed for the mean vitamin C square root
transformed and the percentage of persons with vitamin C deficiency. The CIs for mean vitamin C square root transformed were
constructed on the square root scale with Walds method (34) and
then back-transformed. Because the percentage with vitamin C
,11.4 lmol/L in most subgroups was relatively small, ranging
from 1.0% to 18.0% for NHANES 20032004 and from 1.8% to
31.3% for NHANES III, CIs for these percentages were constructed with the arc-sine transformation (32). NHANES III analytic guidelines (23) were used to assess the stability of the
percentiles. Minimum sample size needed to present estimated
percentiles are a function of the design effect that measures the
effect of the complex sample design on the variance estimate and
is defined as the ratio of the design-based variance to the variance
of a simple random sample of the same size.
Means of vitamin C concentrations (square root transformed)
and percentage of persons with vitamin C concentrations , 11.4
lmol/L for those aged 6 y and those aged 20 y were age
adjusted by the direct method with the use of the projected US
Census population estimates from the year 2000 (35). No discernible bias in serum vitamin C was found because of nonresponse on the basis of age, sex, race-ethnicity, cotinine, BMI,
vitamin C supplement use, or dietary intake of vitamin C in
either survey period (data not shown).
The data are presented 1) by sex cross-classified by age
groups (611 y, 1219 y, 2039 y, 4059 y, and 60 y) and 2) by
sex cross-classified by race-ethnicity (non-Hispanic white, non-

Downloaded from www.ajcn.org by on November 6, 2009

Adiposity of adults 20 y was assessed with the use of body

mass index (BMI; kg/m2). In both surveys, height and weight
were measured by trained interviewers with the use of standardized protocols and calibrated equipment. Adults were
classified as being obese (BMI 30), overweight (BMI of 25
29.9), or healthy weight (BMI of 18.524.9) (29). Pregnant females, as determined by a combination of self-report and urinary
chorionic gonadotropin, were excluded from the BMI data
analysis. Results for underweight adults (BMI ,18.5) were not
presented because the sample sizes were too small to produce
statistically reliable results.

VITAMIN C DISTRIBUTION IN THE UNITED STATES

RESULTS

Age and sex

Marked age-related differences in mean concentrations of
vitamin C were evident in the US population and showed
a significant quadratic trend in males (P , 0.001) and females
(P , 0.001) (Table 1). Boys 611 y of age had the highest mean
serum concentrations of any male age group. Similarly, girls
611 y of age had the highest mean serum concentrations of any
female age group. Both sexes showed a decline in mean concentrations during adolescence (P , 0.001). In males, mean
values continued to decline from 1219 to 2029 y (P , 0.001)

to a plateau between the ages of 20 and 59 y (P = 0.546) and

then increased with advancing age (P , 0.001). In females, the
age pattern was somewhat different. Serum vitamin C decreased
linearly from a high point at 611 y of age to a low at 2039 y of
age (P , 0.001) and then increased to 60 y of age. Females
12 y of age had significantly higher mean concentrations of
vitamin C than did their male counterparts. Selected percentile
data (5th, 10th, 25th, 50th, 75th, 90th, and 95th) for each age
group are presented in Table 1.
Race-ethnicity and sex
With the comparison of sexes, women had higher mean
concentrations of vitamin C than did men in each of the 3 raceethnic groups (Table 1). Within sex but stratified by race-ethnic
group, no significant differences were observed in mean concentrations of vitamin C among men of different race-ethnicity,
but among women of different race-ethnicity, non-Hispanic white
women had significantly higher mean concentrations than did
non-Hispanic black women. Selected percentile data for each sex
stratified by race-ethnicity are shown in Table 1.
Smoking
The mean serum concentration of vitamin C of all smokers was
33% lower than that of all nonsmokers (data not shown). Men
who smoked had mean serum concentrations of vitamin C that
were 31% lower than nonsmoking men, whereas women smokers
had concentrations that were 33% lower than nonsmoking
women (Table 2). Selected percentile data for each smoking
status group stratified by sex are shown in Table 2.
Body mass index
Mean serum vitamin C was significantly lower in obese men than
in overweight men (15% lower) but not significantly different
between overweight and healthy-weight men (Table 2). For women,
mean vitamin C was significantly lower in obese than in overweight
(15% lower) or healthy-weight (25% lower) women. Selected
percentile data for each BMI category are shown in Table 2.
Socioeconomic status
In men and in women, mean vitamin C increased linearly with
increasing PIR (Table 2). In men, vitamin C concentrations were
significantly higher in high compared with low PIR groups (33%
higher). In women, vitamin C concentrations were significantly
higher in medium compared with low PIR (20% higher), high
compared with medium PIR (14% higher), and high compared
with low PIR (37% higher) groups. Selected percentile data for
each PIR category are shown in Table 2.
Vitamin C supplement use
Mean serum concentrations of vitamin C in different age
groups in vitamin C supplement users and nonusers are shown in
Figure 1. The quadratic age trends in mean concentrations of
vitamin C seen in all males and females were retained when the
data were stratified by supplement usage (P for trend , 0.001).
Among nonusers, boys had significantly higher means than did
girls (P , 0.01); however, in age groups 12 y, female nonusers

Downloaded from www.ajcn.org by on November 6, 2009

Hispanic black, Mexican American), smoking status (smoker,

nonsmoker), and BMI category (healthy weight, overweight,
obese), socioeconomic status (low, medium, high), vitamin C
supplement use (any, none), and 1-d dietary intake of vitamin C
compared with the EAR (less than EAR or greater than or equal
to EAR) for both survey periods (19881994, 20032004). To
determine whether stratification was needed in the analysis of
mean vitamin C and the percentage of persons with deficient
concentrations of vitamin C, we tested for the presence of
2-factor interactions of sex with age group, race-ethnicity, smoking status, BMI, income, vitamin C supplement use, and dietary
intake. For NHANES 20032004, we showed the existence of
interactions between sex and age group for mean serum concentrations of vitamin C (P , 0.001) and percentage of persons
who were vitamin C deficient (P , 0.01). An interaction was
also observed between sex and race-ethnicity for mean concentrations of vitamin C (P = 0.046). No other 2-factor interactions were significant (P . 0.05).
Equality of means or equality of percentages was tested
univariately at the a = 0.05 level with the Students t statistic
(33). The equality of .2 subgroups was tested simultaneously.
If the hypothesis that the means or percentages of all subgroups
were equal was rejected, pairwise tests were performed applying
the Bonferroni method (36) to control for multiple comparisons.
To test for linear and quadratic trends in age and income level,
the null hypothesis of no linear or quadratic trend was examined
with orthogonal contrast matrices (37). Rejection of this hypothesis implied the existence of a linear or quadratic trend. To
investigate the odds ratio for vitamin C deficiency in smokers
and nonsmokers during 2 survey periods while controlling for
possible confounding variables, a multiple logistic regression
analysis was performed. Odds ratios having a 95% CI not including unity were considered significant. We investigated the
odds ratios of supplement users and nonusers in a similar
manner. The hypothesis that the odds ratio was equal to unity
was tested by testing the equivalent hypothesis that the log of the
odds ratio was equal to 0.
SAS 9.1 (SAS Institute, Cary, NC) and SUDAAN 10.0 (RTI,
Research Triangle Park, NC) were used to construct CIs for the
means and percentages, to model odds ratios for vitamin C
deficiency among smokers and vitamin C supplement users
(SUDAAN, PROC RLOGIST), and to test statistical hypotheses.
Unless specified otherwise, data are presented as backtransformed weighted square root transformed mean with 95%
CI. Serum vitamin C is indicated in lmol/L. To convert lmol/L
to mg/dL, multiply lmol/L by 0.0176.

1255

1256

SCHLEICHER ET AL

TABLE 1
Serum vitamin C concentrations (in lmol/L) of persons in different age or race-ethnic groups, stratified by sex in the United States, 20032004
95% Confidence
limits
Group

Mean1

7277
4438

Lower

Upper

5th

10th

25th

50th

75th

90th

95th

51.44
49.05

48.4
45.8

54.6
52.3

8.5
7.6

14.7
13.1

34.5
31.7

56.3
54.4

72.8
70.7

90.7
88.9

103.5
102.1

3590
2153
400
1037
725
628
800

48.04
44.85
73.5
50.77
42.07
43.27
52.57

44.9
41.3
69.8
47.2
37.6
38.9
48.8

51.2
48.4
77.3
54.2
46.6
47.6
56.4

7.3
6.6
6
13.6
6.7
6.2
8.1

13.1
10.7
43.6
20.6
10.3
9.5
15.0

31.0
27.1
59.8
36.8
26.0
25.5
35.6

52.7
50.1
76.7
54.3
46.6
48.7
57.7

68.3
65.5
88.0
67.9
62.3
64.0
74.3

85.5
80.3
108.0
85.1
75.6
79.6
97.6

100.4
98.2
6
92.4
87.0
98.1
113.5

3687
2285
423
979
815
638
832

54.84
53.15
68.9
54.88
48.88
52.08
62.98

51.6
49.9
64.7
50.6
44.5
47.5
60.5

58.0
56.5
73.3
59.2
53.3
56.7
65.4

9.8
9.2
6
13.5
8.4
7.9
13.1

16.7
15.1
36.4
19.0
13.3
15.0
21.8

38.7
37.5
55.5
38.2
33.0
36.3
47.7

59.8
58.9
70.0
59.2
54.1
57.8
66.8

76.9
75.8
87.2
75.9
71.7
73.1
86.1

94.8
93.1
106.9
91.6
87.4
89.8
106.7

107.6
107.0
6
102.3
97.7
102.6
120.8

433
1149
410

44.35
44.65
42.45

39.5
40.5
38.5

49.4
49.0
46.5

6
6.4
6.8

18.0
9.5
12.8

36.4
25.7
27.9

51.6
50.8
46.4

62.0
66.8
62.2

72.5
83.7
76.6

6
101.1
88.0

459
1217
441

51.25,10
54.15,10
46.35,11

45.7
49.7
43.4

57.0
58.8
49.4

6
8.5
10.6

22.7
14.1
16.1

39.7
38.5
31.3

54.6
61.0
49.0

70.0
78.6
65.2

82.6
96.6
80.7

6
112.1
89.8

Weighted square root transformed; equality of means tested on a square root scale with Students t statistic with 15 df.
Calculated on the basis of raw data (weighted untransformed data).
3
Includes all race-ethnic categories.
4
Age-adjusted by using the direct method to the year 2000 US census population estimates with the use of age groups 611, 1219, 2039, 4059, and
60 y (35).
5
Adjusted by using the direct method to the year 2000 US census population estimates with the use of age groups 2039, 4959, and 60 y (35).
6
Percentiles do not meet standards for reliability because of small cell size (23).
7
Significantly different from 611-y-old boys, P , 0.001.
8
Significantly different from 611-y-old girls, P , 0.001, and males in respective age groups, P , 0.005.
9
Race and ethnicity were self-reported.
10
Significantly different from men in respective race-ethnic group, P , 0.001.
11
Significantly different from non-Hispanic black men, P , 0.05, and non-Hispanic white women, P , 0.001.
2

had significantly higher mean concentrations than did male

nonusers (Figure 1). Among users, women aged 60 y had
significantly higher means than did men aged 60 y older (P ,
0.001). In those adults aged 20 y for whom serum vitamin C
measurements were available, 37% of men and 47% of women
reported consuming one or more vitamin Ccontaining supplement in the past 30 d (Table 3). Vitamin C supplement users had
significantly higher mean serum concentrations of vitamin C
than did nonusers of vitamin C supplements. Selected percentiles of serum concentrations of vitamin C are presented by sex
and use of vitamin C supplements in Table 3.
Dietary intake of vitamin C
Among adults aged 20 y who had serum measurements of
vitamin C available, 60% of adult men and 53% of adult women

reported dietary intake of vitamin C less than the EAR. Men and
women with vitamin C intakes greater than or equal to the EAR
on the day before examination had significantly higher mean
concentrations of vitamin C than did their counterparts whose
vitamin C intake was less than the EAR (Table 3). Within each
vitamin C dietary intake group, mean serum vitamin C concentrations of women were higher than those of men.
Vitamin C deficiency in NHANES 20032004
A serum concentration ,11.4 lmol/L is considered to be
indicative of vitamin C deficiency at which time clinical features
of manifest scurvy may be seen (38). Of the total population
in NHANES 20032004, 7.1 6 0.9% (6SE) were deficient
(Table 4). Only a small percentage of 611-y-old participants
(,2%) and relatively few adolescents (,4%) were deficient. A

Downloaded from www.ajcn.org by on November 6, 2009

Age adjusted3
6 y
20 y
Males3
6 y
20 y
611 y
1219 y
2039 y
4059 y
60 y
Females3
6 y
20 y
611 y
1219 y
2039 y
4059 y
60 y
Adults age 20 y by race-ethnicity9
Men
Mexican American
Non-Hispanic white
Non-Hispanic black
Women
Mexican American
Non-Hispanic white
Non-Hispanic black

Percentile2

1257

VITAMIN C DISTRIBUTION IN THE UNITED STATES

TABLE 2
Serum vitamin C concentrations (in lmol/L) of adults 20 y of all race-ethnic categories in different smoking, BMI, or socioeconomic status categories,
stratified by sex in the United States, 20032004
95% Confidence
limits
Group

Percentile2

Mean1

Lower

Upper

5th

10th

25th

50th

75th

90th

95th

1407
744

50.74
35.2

47.2
32.5

54.3
38.0

10.3
5.3

18.3
7.1

37.5
16.4

55.5
37.2

68.7
57.5

83.4
72.3

101.0
86.8

1844
439

58.04
38.6

54.9
35.5

61.2
41.9

13.4
4.1

23.3
7.6

44.2
16.5

62.4
41.8

78.5
61.7

95.5
80.2

110.3
99.9

597
857
632

46.06
47.07
40.08

40.5
43.3
35.3

51.7
50.9
45.1

6.7
7.5
6.2

10.4
13.7
9.1

26.4
30.7
23.8

52.8
53.6
44.4

68.4
66.5
60.8

86.2
81.4
76.2

100.8
98.1
88.1

636
635
742

60.3
52.89
45.010

57.4
49.2
41.6

63.3
56.6
48.6

9.3
10.6
8.6

16.8
20.4
13.0

46.8
38.6
26.8

65.1
59.5
49.8

80.4
76.1
66.9

99.5
91.7
87.1

115.5
107.4
99.8

329
900
810

36.1
42.4
48.114

30.9
39.5
44.6

41.7
45.5
51.7

13
6.4
7.4

6.7
10.2
13.5

17.5
24.0
31.7

43.3
47.4
52.4

59.2
64.0
67.3

71.8
76.9
85.5

13
87.7
101.6

443
948
765

42.8
51.215
58.516

38.4
47.7
55.7

47.4
54.7
61.4

6.5
9.0
10.7

10.0
14.2
21.9

23.6
35.4
44.3

47.8
58.5
62.9

64.3
76.9
78.6

80.4
95.1
96.1

93.0
107.1
114.7

Weighted square root transformed; age-adjusted with the direct method to the year 2000 US census population with the use of the age groups 2039, 40
59, and 60 y (35); equality of means tested on a square root scale with Students t statistic with 15 df.
2
Calculated on the basis of raw data (weighted untransformed data).
3
Smoker is defined as having serum cotinine . 10 ng/mL; nonsmoker has cotinine  10 ng/mL.
4
Significantly different from smokers, P , 0.001.
5
Rounded to the nearest 10th; pregnant women and adults with BMI (in kg/m2) ,18.5 were excluded. BMI categories are defined as 18.5
to ,25 (healthy weight), 25 to ,30 (overweight), and 30 (obese).
6,10
Significantly different from healthy-weight women: 6P , 0.001, 10P , 0.0001.
7
Significantly different from overweight women, P , 0.001.
8
Significantly different from overweight men, P , 0.01.
9
Significantly different from obese women, P , 0.001.
11
Defined as poverty-income ratio (PIR) of ,1 (low), 1 to ,3 (medium), or 3 (high).
12
Means increased linearly with increasing PIR, P , 0.05 (Students t statistic with 15 df).
13
Percentiles do not meet standards for reliability because of small cell size (23).
14
Significantly different from men with low PIR, P , 0.01.
15
Significantly different from women with high PIR and with low PIR, P , 0.05.
16
Significantly different from women with low PIR, P , 0.05.

significant quadratic age trend in prevalence of deficiency existed in males (P , 0.01) and in females (P , 0.01). Men aged
2039 and 60 y had a significantly higher prevalence of vitamin C deficiency than did women in these age groups (Table
4). The prevalence of deficiency did not differ by race-ethnic
group; however, men in each race-ethnic group were significantly more likely to be vitamin C deficient than were women in
the same race-ethnic group (Table 4).
The percentage of adults with deficient concentrations of vitamin C in NHANES 20032004 was markedly higher among
smokers than among nonsmokers (Table 5). Smokers were at
risk of deficiency .3 times as often as nonsmokers. In adults,
BMI was not related to the prevalence of vitamin C deficiency

(Table 5). The percentage of men and women with vitamin C

deficiency decreased linearly with increasing PIR (Table 5). The
prevalence of vitamin C deficiency was higher in low-income
(17.4%) compared with the high-income (7.9%) men and in
low-income (10.4%) compared with high-income (5.0%)
women.
Adults who were nonusers of vitamin C supplements had
a significantly higher prevalence of vitamin C deficiency than did
users (Table 6). The prevalence of vitamin C deficiency among
nonusers of vitamin C supplements was higher in men and in
women. Adults with vitamin C intake less than the EAR had
a significantly higher prevalence of vitamin C deficiency than
did adults with vitamin C intake greater than or equal to the

Downloaded from www.ajcn.org by on November 6, 2009

Smoking
Men
Nonsmokers
Smokers
Women
Nonsmokers
Smokers
BMI5
Men
Healthy weight
Overweight
Obese
Women
Healthy weight
Overweight
Obese
Socioeconomic status11
Men12
Low
Medium
High
Women12
Low
Medium
High

1258

SCHLEICHER ET AL

EAR. Among adults with vitamin C intakes greater than or equal

to the EAR, the prevalence of vitamin C deficiency was higher
in men than in women (Table 6).
Serum concentrations of vitamin C ,20 lmol/L may be associated with fatigue and irritability (40). In NHANES 2003
2004, the age-adjusted prevalence (6SE) of persons 6 y of age
with concentrations below this cutoff was 13.7 6 1.5%, whereas
the age-adjusted prevalence (6SE) of adults 20 y with concentrations below this cutoff was 15.7 6 1.6%. Persons with
serum vitamin C concentrations , 28 lmol/L are at moderate
risk of developing vitamin C deficiency (41). The age-adjusted
prevalence (6SE) of persons 6 y of age with concentrations ,
28 lmol/L was 19.6 6 1.9%, whereas the age-adjusted prevalence of adults 20 y with concentrations below this cutoff was
22.2 6 2.1%.
Comparison of vitamin C status in NHANES 20032004
and 19881994
Comparison of NHANES 20032004 with NHANES III
(19881994) was undertaken to determine whether serum vitamin C deficiency in the US population had significantly changed.
In NHANES 20032004, the overall prevalence decreased 44%.
The age-related prevalence of deficiency rates for the 2 surveys
was parallel; deficiency rates were low in children, climbed to
peak at middle age, and fell in the older age groups (quadratic
trend in both males and females, P , 0.001).
The prevalence of vitamin C deficiency in all adults aged
20 y stratified by race-ethnicity in NHANES III was 14.7% in
non-Hispanic whites, 21.9% in non-Hispanic blacks, and 14.2%
in Mexican Americans; these are in contrast to lower prevalence
rates of 10.0%, 6.7%, and 5.9%, respectively in 20032004.
When stratified by sex, vitamin C deficiency significantly decreased in all 3 race-ethnic groups [non-Hispanic blacks (67
73% decrease), Mexican Americans (5662% decrease), and
non-Hispanic whites (3233% decrease); Table 4].

DISCUSSION

Data on serum vitamin C from the nationally representative

NHANES 20032004 survey show that the vitamin C status
improved in the US population since 19881994. In virtually all
subgroups explored in this analysis, fewer persons were categorized as deficient, and most differences were significant. The
race-ethnic differences in the prevalence of vitamin C deficiency
seen in NHANES III were no longer apparent in the 20032004
survey period. The magnitude of the improvement (5673%) was
striking for minority groups. Although non-Hispanic black
women had significantly lower mean concentrations than did
non-Hispanic white women in NHANES 20032004, the prevalence of deficiency was not lower, a suggestion that there are
different distributions of values in the 2 groups of women. Each
of the 3 race-ethnic groups showed a significant sex difference in
serum concentrations of vitamin C in NHANES 20032004.
Higher concentrations in females have been observed in other
studies in the United States (20, 42) and Europe (4345), although not noted in an Asian population (46). The inverse relation
between BMI and serum concentrations of vitamin C seen in
NHANES 20032004 had been noted previously (45, 47). Lower
vitamin C intake and more prevalent vitamin C deficiency among
persons with low income status have been reported in other
populations (4850).

Downloaded from www.ajcn.org by on November 6, 2009

FIGURE 1. Mean (695% CI) serum concentrations of vitamin C in

selected age groups were stratified by sex and any recent (past 30 d)
intake of vitamin Ccontaining supplements in the US population during
NHANES 20032004. A square root transformation was used to estimate
means. Significant quadratic age trends in mean concentrations of vitamin C
were present in all 4 groups (P , 0.001). Sample sizes for supplement users
and nonusers were as follows: female users (n = 1363; filled square with
solid line); male users (n = 1034; filled diamond with solid line); female
nonusers (n = 2317; open square with dashed line); male nonusers (n = 2550;
open diamond with dashed line).

In both surveys, smokers had a higher prevalence of vitamin C

deficiency than did nonsmokers (Table 5). However, in NHANES
20032004 the prevalence of vitamin C deficiency among all
smokers declined significantly compared with NHANES III (41
42% decrease); nonsmokers showed a similar improvement (42
44% decrease). The significant association of smoking status and
vitamin C deficiency persisted after controlling for the possible
confounding effects of sex, age, race-ethnicity, BMI, income, use
of vitamin Ccontaining supplements, dietary intake of vitamin C,
and survey (odds ratio: 3.76; 95% CI: 3.08, 4.59; P , 0.001).
Despite a decline, smokers remained the subgroup most at risk
of vitamin C deficiency (Table 5). Significant declines in the
prevalence of vitamin C deficiency in 20032004 were observed
in most BMI-related categories except for obese men who had
the smallest improvement in vitamin C deficiency (22% decrease)
of any subgroup (Table 5). In low and medium income but not
high income groups, the prevalence of vitamin C deficiency declined between surveys in all 3 income groups (3449%; Table 5).
Comparing the prevalence of vitamin C deficiency in the 2
surveys by vitamin C supplement use, an improvement was
observed within each of the 4 sex-by-vitamin C supplement use
groups (Table 6). The low prevalence of deficiency in vitamin C
supplement users in NHANES III was further reduced by 60
62% in NHANES 20032004. The significant association of
vitamin C supplement use and low risk of vitamin C deficiency
persisted after controlling for the possible confounding effects of
sex, age, race-ethnicity, BMI, PIR, smoking status, dietary intake
of vitamin C, and survey (odds ratio: 5.21; 95% CI: 3.98, 6.83;
P , 0.001). In both categories of dietary intake, vitamin C status
improved in NHANES 20032004 compared with NHANES III
(Table 6). Women whose dietary intake of vitamin C was greater
than or equal to the EAR were the least likely of any subgroup to
be vitamin C deficient and showed the greatest improvement in
status since NHANES III (80% decrease in deficiency).

1259

VITAMIN C DISTRIBUTION IN THE UNITED STATES

TABLE 3
Serum vitamin C concentrations (in lmol/L) of adults 20 y for all race-ethnic groups in different categories of supplement use or dietary intake, stratified
by sex in the United States, 200320041
95% Confidence
limits
Group

Percentile3

Mean2

Lower

Upper

5th

10th

25th

50th

75th

90th

95th

788
1360

60.65
35.1

58.1
31.4

63.1
39.1

19.7
5.2

32.9
7.9

48.7
18.6

61.7
38.5

76.2
57.1

98.2
68.6

111.4
75.9

1,070
1,211

66.36
42.27

63.3
38.5

69.4
46.0

26.0
6.9

37.9
10.0

54.5
23.7

67.7
46.6

84.6
65.2

102.7
83.4

120.4
92.7

816
1227

58.69
37.6

53.9
33.9

63.4
41.4

1011
1159

64.310
46.111

61.1
42.3

67.6
50.1

Vitamin C supplement use

Men
Users
Nonusers
Women
Users
Nonusers
Vitamin C dietary intake8
Men
Greater than or equal to the EAR
Less than the EAR
Women
Greater than or equal to the EAR
Less than the EAR
1

On the basis of NHANES serum cotinine data, the prevalence

of adult smoking decreased from 25.5 to 22.7% between 1990
and 1999 (51). Serum cotinine concentrations in nonsmokers
decreased by 70% between 1988 and 2002 (52). Among
smokers, there has been a decrease in cigarette use. From
NHANES III to NHANES 19992000, the average number of
cigarettes smoked per day fell by 15%, and mean serum cotinine
fell in smokers by 13% (53). Since NHANES III, many communities have banned smoking in public places and consequently reduced passive exposure to secondhand smoke (52).
Thus, it seems likely that a reduction in smoke exposure in
smokers and nonsmokers alike may account in part for the improvement in vitamin C concentrations in the population during
20032004.
The relation between serum concentration of vitamin C and
age in the US population is complex and nonlinear. This relation
appears to be different from country to country, depending on
such factors as the availability of fruit and vegetables, socioeconomic levels, and supplement usage (48, 54). Fruit and
vegetable consumption among American adults remained relatively stable from 1994 through 2005 (55). For example, average
daily fruit intake in persons 2 y of age remained the same from
19941996 to 19992002 (1.6 servings), and average vegetable
consumption declined slightly from 3.4 to 3.2 servings/d during
the same period (56). Increased intake of vitamin Ccontaining

foods was unlikely to have contributed to the reduced prevalence

of vitamin C deficiency during the recent survey. Mean 1-d intakes of vitamin C from food in those aged 12 y ranged from
91 to 125 mg during NHANES III (20), whereas in NHANES
20032004, mean 1-d intakes were slightly lower, ranging from
80 to 116 mg (57). In our analysis of NHANES 20032004,
more than one-half the adults who had vitamin C measurements
available had a 1-d dietary intake less than the EAR, and those
with intake less than the EAR had serum concentrations of vitamin C that were approximately one-third lower than those with
adequate dietary intake, and they were 510 times more likely to
be vitamin C deficient. It should be noted that the dietary intake
data were from a single 24-h dietary recall and thus may not
provide the best estimate of usual dietary intake.
In prosperous societies, supplement consumption has a significant effect on body stores and circulating concentrations of
vitamin C. In NHANES 19992000, 52% of adults reported
consumption of supplements in the past month, and 35% of adults
were regular users of multivitamins (58). Usage rates in children
were similar but lower in adolescents. These recent data show
increased usage since the overall 40% usage reported during
NHANES III (58) and are likely to explain in part the improved
vitamin C status of the US population. In NHANES 20032004
for those indicating vitamin C supplement use during the preceding month, serum concentrations were 5773% higher than in

Downloaded from www.ajcn.org by on November 6, 2009

EAR, Estimated Average Requirement.

Weighted square root transformed; age-adjusted with the direct method to the year 2000 US census population with the use of the age groups 2039, 40
59, and 60 y (35); equality of means tested on a square root scale with Students t statistic with 15 df.
3
Calculated on the basis of raw data (weighted untransformed data); data not shown for categories based on 1-d estimate of dietary intake because of
overestimation of probabilities in tails (31).
4
Defined as having taken one or more vitamin Ccontaining supplement during the previous 30 d.
5
Significantly different from men who were nonusers and women who were users, P , 0.001 (Students t statistic with 15 df).
6
Significantly different from women who were nonusers, P , 0.001 (Students t statistic with 15 df).
7
Significantly different from men who were nonusers, P , 0.001 (Students t statistic with 15 df).
8
Based on a single 24-h recall interview; EAR is defined as the daily intake that is estimated to meet the requirement in one-half of apparently healthy
persons in a life-stage or sex group (31).
9
Significantly different from men with intake less than the EAR and women with intake greater than or equal to the EAR, P , 0.001 (Students t statistic
with 15 df).
10
Significantly different from women with intake less than the EAR, P , 0.001 (Students t statistic with 15 df).
11
Significantly different from men with intake less than the EAR, P , 0.001 (Students t statistic with 15 df).
2

1260

SCHLEICHER ET AL

TABLE 4
Serum vitamin C deficiency (,11.4 lmol/L) of persons in different age or race-ethnic groups, stratified by sex, with upper (UL) and lower (LL) 95%
confidence limits in the United States, for 19881994 and 200320041
20032004
Group

Percentage

LL

UL

Percentage

LL

UL

P value2

7277
4438

7.1
8.4

5.3
6.2

9.2
10.9

20,636
15,185

12.8
14.8

11.1
13.0

14.5
16.7

,0.001
,0.001

3590
2153
400
1037
725
628
800

8.2
10.0
1.36
2.77
10.88
11.0
7.28

5.7
6.9
0.1
1.1
7.4
6.5
4.5

11.2
13.6
3.6
5.1
14.7
16.4
10.3

9774
7123
1347
1304
2823
1921
2379

15.1
17.6
2.07
9.8
17.7
19.7
14.0

13.3
15.6
0.8
6.5
15.2
16.5
12.2

17.0
19.8
3.7
13.6
20.5
23.1
16.0

0.001
,0.001
0.544
0.002
0.008
0.012
0.002

3687
2285
423
979
815
638
832

6.0
6.9
1.86
3.9
7.9
7.6
4.1

4.4
5.1
0.3
1.9
5.1
5.0
2.9

7.8
8.9
4.6
6.5
11.1
10.6
5.6

10,862
8062
1290
1510
3398
2179
2485

10.6
12.2
1.87
7.6
14.9
11.8
8.3

9.0
10.4
0.7
4.8
12.6
9.4
6.5

12.4
14.2
3.4
10.9
17.4
14.3
10.2

0.001
,0.001
0.992
0.064
0.002
0.037
0.002

1163
416
433

11.810
8.910
7.710

8.0
5.6
3.7

16.1
12.8
13.0

2976
1838
2052

17.4
26.9
17.4

15.0
24.6
14.7

20.1
29.3
20.2

0.022
,0.001
0.001

1231
448
459

8.2
5.0
4.27

6.0
3.0
1.8

10.8
7.4
7.5

3408
2266
2038

12.2
18.2
11.1

9.9
16.2
8.9

14.7
20.2
13.5

0.021
,0.001
,0.001

Confidence limits were constructed with the arc-sine transformation (32). Data for 19881994 were based on the third National Health and Nutrition
Examination Survey (NHANES).
2
Equality of percentages between survey periods (20032004 and 19881994) were tested on the arc-sine scale by using the Students t statistic with 15 df.
3
Includes all race-ethnic categories.
4
Percentages were age-adjusted by using the direct method to the year 2000 US census population estimates with the use of age groups 611, 1219, 20
39, 4059, and 60 y (35).
5
Percentages were age-adjusted by using the direct method to the year 2000 US census population estimates with the use of age groups 2039, 4959 y,
and 60 y (35).
6
Relative SE = (SE of the percentage/percentage) 100 .40% (39).
7
Relative SE = (SE of the percentage/percentage) 100 .30% but 40% (39).
8
Significantly different from females in respective age group, P , 0.05 (Students t statistic with 15 df).
9
Race and ethnicity were self-reported.
10
Significantly different from women in respective race-ethnic group, P , 0.05 (Students t statistic with 15 df).

nonusers. When the relation between age and vitamin C status

was examined in supplement users and nonusers, older persons
(60 y) had better vitamin C status in part because of supplement use. A frequency analysis of smoking by supplement usage
in this population showed that those aged 60 y who did not use
supplements were more likely to be smokers than those who did
(22% compared with 12%). Smoking accelerates vitamin C
turnover and lowers serum concentrations. Although older
women are more likely to consume fruit and vegetables than
their younger counterparts, accounting for some of the improvement in the oldest age group, older men are not (59). Even
so, a substantial minority (10%) of older men and women
consume far less vitamin C than the Recommended Dietary
Allowance through dietary sources (59).
The consequences of vitamin C deficiency in adults range from
petechial hemorrhages, follicular hyperkeratoses with unerupted

corkscrew hairs, anemia, bleeding gums, and loosened teeth to

more subtle changes in mood and affect which might become
apparent before the body is fully depleted. Mild but distinct
fatigue and irritability in the absence of manifest scurvy but
consistent with latent scurvy were reported by 6 of 7 healthy
young men subjected to a vitamin C depletion protocol (60).
Latent scurvy is also characterized by nonspecific symptoms such
as joint and muscle pain (3). A substantial percentage of adults
(16%) in NHANES 20032004 had vitamin C concentrations that
are associated with low energy and weakness as a result of inadequate intake of vitamin C. More than 20% of adults showed
marginal vitamin C status, placing them at risk of vitamin C
deficiency, similar to the 2023% estimates in NHANES III (20).
The limitations of the study are several. For the survey
comparison, NHANES III data quality was not as good as
NHANES 20032004. Although the main indicators of data

Downloaded from www.ajcn.org by on November 6, 2009

Age adjusted3
6 y4
20 y5
Males3
6 y4
20 y5
611 y
1219 y
2039 y
4059 y
60 y
Females3
6 y4
20 y5
611 y
1219 y
2039 y
4059 y
60 y
20 y and race-ethnicity5,9
Men
Non-Hispanic white
Non-Hispanic black
Mexican American
Women
Non-Hispanic white
Non-Hispanic black
Mexican American

19881994

1261

VITAMIN C DISTRIBUTION IN THE UNITED STATES

TABLE 5
Serum vitamin C deficiency (,11.4 lmol/L) of persons aged 20 y in different categories of smoking, BMI, or socioeconomic status, stratified by sex, with
upper (UL) and lower (LL) 95% confidence limits in the United States, for 19881994 and 200320041
20032004
Group

Percentage

LL

UL

Percentage

LL

UL

P value2

1407
744

5.34
18.0

2.9
13.6

8.3
23.0

4468
2604

9.5
30.4

8.0
26.7

11.2
34.3

0.010
,0.001

1844
439

4.24
15.3

2.7
12.0

6.0
18.9

6044
1937

7.2
25.9

5.8
21.7

8.7
30.3

0.008
,0.001

597
857
632

11.1
7.6
12.7

6.4
5.3
8.0

16.9
10.3
18.3

2672
2896
1449

18.1
17.3
16.2

15.6
14.5
12.8

20.7
20.4
19.8

0.017
,0.001
0.135

636
635
742

6.7
5.7
8.1

5.0
3.5
4.9

8.7
8.3
12.0

2870
2369
2349

10.6
11.1
17.5

8.5
8.0
14.7

12.9
14.6
20.4

0.007
0.007
,0.001

329
900
810

17.4
10.9
7.98

12.5
6.6
4.7

22.9
16.1
11.9

1291
3093
2117

31.3
21.2
12.0

26.8
18.3
9.6

35.9
24.3
14.7

,0.001
0.001
0.042

443
948
765

10.4
7.7
5.08

6.9
5.5
3.2

14.7
10.2
7.1

1868
3360
2050

20.1
14.4
7.7

16.0
12.4
5.5

24.6
16.5
10.1

0.002
,0.001
0.044

Includes all race-ethnic categories; age-adjusted by the direct method to the year 2000 US census population estimates with the use of age groups 2039,
4959, and 60 y (35). Confidence limits were constructed with the arc-sine transformation (32). Data for 19881994 were based on data from the third
National Health and Nutrition Examination Survey (NHANES).
2
Equality of percentages between survey periods (20032004 and 19881994) were tested on the arc-sine scale (Students t statistic with 15 df).
3
Smoker is defined as having serum cotinine . 10 ng/mL; nonsmoker has cotinine 10 ng/mL.
4
Significantly different from smokers in respective sex groups, P , 0.05 (Students t statistic with 15 df).
5
Rounded to the nearest 10th; pregnant women and adults with BMI (in kg/m2) ,18.5 were excluded. BMI categories are defined as 18.5 to ,25
(healthy weight), 25 to ,30 (overweight), and 30 (obese).
6
Defined as poverty-income ratio (PIR) of ,1 (low), 1 to ,3 (medium), or 3 (high).
7
NHANES 20032004: percentage deficient decreased linearly with increasing PIR, P , 0.01 (Students t statistic with 15 df).
8
Significantly different from low PIR, P , 0.01 (Students t statistic with 15 df).

quality (bench QC pools) were acceptable, the blind QC pools

and split sample data were not optimal. For this reason a more
extensive analysis of the mean concentrations of vitamin C in
various subgroups was not undertaken, but rather we limited our
comparison to prevalence of deficiency. To compare surveys it
was necessary to adjust the NHANES III data set to make it
comparable to the NHANES 20032004 survey method.
Improvements in the assay were made in the period between the 2
surveys. Consumable supplies were shown to contain materials
that measurably degrade vitamin C within a short period of time
(,24 h) (61). It was suggested that this degradation may occur
in autosampler vials, tubes, and vials used in the collection and
processing of blood and serum samples and that this type of
degradation may account for a substantial amount of the observed interlaboratory variation in NIST-sponsored exercises for
vitamin C measurement. Two features of the assay used for
NHANES 20032004 were designed to mitigate effects of oxidation and degradation, namely, use of an internal standard that
partially compensated for procedural losses during the assay and
calibrators that were processed the same as samples such that

both were exposed to similar conditions during the processing

and analysis steps. Overall, 3% of the difference in mean
serum concentrations between survey periods can be explained
by method bias. This amount of bias is considered clinically acceptable, considering other sources of error in measuring serum
vitamin C such as analytic error (CVA = 59%), within-individual
variation for repeated measurements (CVI = 26%), and variation
of the group or population for this analyte (CVG = 31%) (62, 63).
A complication in interpreting vitamin C serum and dietary
intake data, ordinarily considered the sum of intake from fruit,
vegetables, supplements, and fortified foods, is the availability of
D-ascorbic acid, which is used as a preservative in prepared
foods. The HPLC method used in this analysis (and during
NHANES III) does not distinguish between the D- and Lisomers, which probably do not share full biological activity (1).
A significant amount of D-ascorbate is found in some prepared
foods, particularly cured meats where it is used to shorten
processing time and improve the color.
The strengths of this study are numerous. These results are
from a national survey to measure vitamin C with the use of

Downloaded from www.ajcn.org by on November 6, 2009

Smoking3
Men
Nonsmokers
Smokers
Women
Nonsmokers
Smokers
BMI5
Men
Healthy weight
Overweight
Obese
Women
Healthy weight
Overweight
Obese
Socioeconomic6
Men7
Low
Medium
High
Women7
Low
Medium
High

19881994

1262

SCHLEICHER ET AL

TABLE 6
Serum vitamin C deficiency (,11.4 lmol/L) of persons aged 20 y in different categories of supplement use or dietary intake, stratified by sex, with upper
(UL) and lower (LL) 95% confidence limits in the United States, 19881994 and 200320041
20032004
Group
Vitamin C supplement use3
Men
Users
Nonusers
Women
Users
Nonusers
Vitamin C dietary intake5
Men
Greater than or equal to the EAR
Less than the EAR
Women
Greater than or equal to the EAR
Less than the EAR

19881994

Percentage

LL

UL

Percentage

LL

UL

P value2

788
1360

2.1
15.54

0.7
10.8

4.3
20.9

1755
5368

5.5
23.0

4.0
20.4

7.2
25.6

0.017
0.019

1070
1211

1.9
11.24

1.1
8.3

2.9
14.5

2537
5525

4.8
17.0

3.4
14.7

6.3
19.4

0.003
0.009

816
1227

3.0
15.06

0.8
10.5

6.7
20.2

3343
3516

7.9
26.8

6.3
23.6

9.6
30.2

0.012
,0.001

1011
1159

1.0
10.86

0.6
7.8

1.6
14.2

4123
3692

5.1
19.7

4.1
16.5

6.2
23.1

,0.001
,0.001

a specific and sensitive HPLC method. Estimated prevalence rates

of deficiency are based on serum concentrations of vitamin C, not
from dietary recall. The quality of the NHANES 20032004
laboratory data were excellent as judged by all relevant quality
assurance indicators. NHANES focuses on noninstitutionalized
persons, whereas nutritional data in the elderly are often limited
to those who are not community dwelling.
In conclusion, the vitamin C status of the US population
appears to have substantially improved from 19881994 to 2003
2004. Nevertheless, the prevalence of vitamin C deficiency in
various subgroups remains a concern, considering the wide
availability of vitamin C in common fruit and vegetables, as well
as in fortified foods and beverages.
We thank the staff of the Nutritional Biomarkers Branch who worked on
vitamin C testing past and present, including Mary Xu who measured serum
vitamin C concentrations and Huiping Chen who reviewed the data for quality
assurance for the NHANES 20032004 survey, and Jaime Gahche for advice
on vitamin C supplements.
The authors responsibilities were as followsRLS and DAL: data review;
MDC: statistical analysis; and RLS and MDC: writing of the manuscript. All
authors interpreted data and critically revised the manuscript. None of the
authors had a conflict of interest.

REFERENCES
1. Sauberlich HE. Ascorbic acid. In: Ziegler EE, Filer JLJ, eds. Present
knowledge in nutrition. Washington, DC: ILSI, 1996:13241.
2. Huang A, Vita JA, Venema RC, Keaney JF Jr. Ascorbic acid enhances
endothelial nitric-oxide synthase activity by increasing intracellular
tetrahydrobiopterin. J Biol Chem 2000;275:17399406.
3. Levine M, Rumsey S, Wang Y, Park J, Kwon W, Amano N. Vitamin C.
In: Ziegler EE, Filer LJ Jr, eds. Present knowledge in nutrition. Washington, DC: ILSI, 1996:14659.
4. Olmedo JM, Yiannias JA, Windgassen EB, Gornet MK. Scurvy: a disease almost forgotten. Int J Dermatol 2006;45:90913.

5. Prinzo ZW. Scurvy and its prevention and control in major emergencies.
Geneva, Switzerland: World Health Organization, 1999. (WHO publication WHO/NHD/99.11.)
6. Block G. Vitamin C and cancer prevention: the epidemiologic evidence
[see comments]. Am J Clin Nutr 1991;53(suppl):270S82S.
7. Block G. Vitamin C status and cancer. Epidemiologic evidence of reduced risk. Ann N Y Acad Sci 1992;669:28090 (discussion 2902).
8. Gonzalez CA, Pera G, Agudo A, et al. Fruit and vegetable intake and the
risk of stomach and oesophagus adenocarcinoma in the European Prospective Investigation into Cancer and Nutrition (EPIC-EURGAST). Int
J Cancer 2006;118:255966.
9. Bjelakovic G, Nikolova D, Simonetti RG, Gluud C. Antioxidant supplements for prevention of gastrointestinal cancers: a systematic review
and meta-analysis. Lancet 2004;364:121928.
10. Gaziano JM, Glynn RJ, Christen WG, et al. Vitamins E and C in the
prevention of prostate and total cancer in men: the Physicians Health
Study II randomized controlled trial. JAMA 2009;301:5262.
11. Lin J, Cook NR, Albert C, et al. Vitamins C and E and beta carotene
supplementation and cancer risk: a randomized controlled trial. J Natl
Cancer Inst 2009;101:1423.
12. Steffen LM, Jacobs DR Jr, Stevens J, Shahar E, Carithers T, Folsom AR.
Associations of whole-grain, refined-grain, and fruit and vegetable
consumption with risks of all-cause mortality and incident coronary
artery disease and ischemic stroke: the Atherosclerosis Risk in Communities (ARIC) Study. Am J Clin Nutr 2003;78:38390.
13. Ness AR, Powles JW. Fruit and vegetables, and cardiovascular disease:
a review. Int J Epidemiol 1997;26:113.
14. Bazzano LA, He J, Ogden LG, et al. Fruit and vegetable intake and risk
of cardiovascular disease in US adults: the first National Health and
Nutrition Examination Survey Epidemiologic Follow-up Study. Am J
Clin Nutr 2002;76:939.
15. Effect of supplemental antioxidants vitamin C, vitamin E, and coenzyme
Q10 for the prevention and treatment of cardiovascular disease. Rockville, MD: Agency for Healthcare Research and Quality, 2003. (Evidence Report/Technology Assessment: No. 83. AHRQ publication
03-E042.)
16. Sesso HD, Buring JE, Christen WG, et al. Vitamins E and C in the
prevention of cardiovascular disease in men: the Physicians Health
Study II randomized controlled trial. JAMA 2008;300:212333.

Downloaded from www.ajcn.org by on November 6, 2009

Includes all race-ethnic categories; age-adjusted by using the direct method to the year 2000 US census population estimates with the use of age groups
2039, 4959, and 60 y (35). Confidence limits were constructed with the arc-sine transformation (32). Data for 19881994 were based on data from
the third National Health and Nutrition Examination Survey (NHANES). EAR, Estimated Average Requirement.
2
Equality of percentages between survey periods (20032004 and 19881994) were tested on the arc-sine scale tested with Students t statistic with 15 df.
3
A supplement user is defined as having taken one or more vitamin Ccontaining supplements during the previous 30 d.
4
Significantly different from vitamin C supplement users in respective sex group, P , 0.05 (Students t statistic with 15 df).
5
Based on a single 24-h recall interview; defined as the daily intake that is estimated to meet the requirement in one-half of apparently healthy persons in
a life-stage or sex group (31).
6
Significantly different from those with vitamin C intake greater than or equal to the EAR in respective sex group, P , 0.05 (Students t statistic with 15 df).

VITAMIN C DISTRIBUTION IN THE UNITED STATES

42. Preston AM, Rodriguez C, Rivera CE. Plasma ascorbate in a population

of children: influence of age, gender, vitamin C intake, BMI and smoke
exposure. P R Health Sci J 2006;25:13742.
43. Faure H, Preziosi P, Roussel AM, et al. Factors influencing blood concentration of retinol, [alpha]-tocopherol, vitamin C, and [beta]-carotene
in the French participants of the SU.VI.MAX trial. Eur J Clin Nutr 2006;
60:70617.
44. Olmedilla B, Granado F, Southon S, et al. Serum concentrations of
carotenoids and vitamins A, E, and C in control subjects from five
European countries. Br J Nutr 2001;85:22738.
45. Galan P, Viteri FE, Bertrais S, et al. Serum concentrations of betacarotene, vitamins C and E, zinc and selenium are influenced by sex,
age, diet, smoking status, alcohol consumption and corpulence in
a general French adult population. Eur J Clin Nutr 2005;59:118190.
46. Benzie IF, Janus ED, Strain JJ. Plasma ascorbate and vitamin E levels in
Hong Kong Chinese. Eur J Clin Nutr 1998;52:44751.
47. Schectman G, Byrd JC, Gruchow HW. The influence of smoking on
vitamin C status in adults [see comments]. Am J Public Health 1989;79:
15862.
48. Kant AK, Graubard BI. Secular trends in the association of socioeconomic position with self-reported dietary attributes and biomarkers
in the US population: National Health and Nutrition Examination Survey (NHANES) 19711975 to NHANES 19992002. Public Health
Nutr 2007;10:15867.
49. Mosdol A, Erens B, Brunner EJ. Estimated prevalence and predictors of
vitamin C deficiency within UKs low-income population. J Public
Health (Oxf) 2008;30:45660.
50. Shohaimi S, Bingham S, Welch A, et al. Occupational social class,
educational level and area deprivation independently predict plasma
ascorbic acid concentration: a cross-sectional population based study in
the Norfolk cohort of the European Prospective Investigation into
Cancer (EPIC-Norfolk). Eur J Clin Nutr 2004;58:14325.
51. Centers for Disease Control and Prevention. State-specific prevalence of
current cigarette smoking among adults, and policies and attitudes about
secondhand smokeUnited States, 2000. MMWR Morb Mortal Wkly
Rep 2001;50:11016.
52. Pirkle JL, Bernert JT, Caudill SP, Sosnoff CS, Pechacek TF. Trends in
the exposure of nonsmokers in the U.S. population to secondhand
smoke: 19882002. Environ Health Perspect 2006;114:8538.
53. OConnor RJ, Giovino GA, Kozlowski LT, et al. Changes in nicotine
intake and cigarette use over time in two nationally representative
cross-sectional samples of smokers. Am J Epidemiol 2006;164:7509.
54. Villalpando S, Montalvo-Velarde I, Zambrano N et al. Vitamins A, and
C and folate status in Mexican children under 12 years and women 12
49 years: a probabilistic national survey. Salud Publica Mex 2003;
45(Suppl 4):S50819.
55. Blanck HM, Gillespie C, Kimmons JE, Seymour JD, Serdula MK.
Trends in fruit and vegetable consumption among U.S. men and women,
19942005. Prev Chronic Dis 2008;5:110.
56. Centers for Disease Control and Prevention (CDC). Fruit and vegetable
consumption among adultsUnited States, 2005. MMWR Morb Mortal
Wkly Rep 2007;56:2137.
57. USDA ARS. Nutrient intakes from food: mean amounts consumed per
individual, one day, 20032004. What we eat in America, NHANES
20032004: documentation and data files (Table 1). 2006. Available
from: http://www.ars.usda.gov/SP2UserFiles/Place/12355000/pdf/0304/
Table_1_NIF.pdf (cited 8 June 2009).
58. Rock CL. Multivitamin-multimineral supplements: who uses them? Am
J Clin Nutr 2007;85(suppl):277S9S.
59. Wakimoto P, Block G. Dietary intake, dietary patterns, and changes with
age: an epidemiological perspective. J Gerontol A Biol Sci Med Sci
2001;56:6580.
60. Levine M, Conry-Cantilena C, Wang Y, et al. Vitamin C pharmacokinetics in healthy volunteers: evidence for a recommended dietary allowance. Proc Natl Acad Sci U S A 1996;93:37049.
61. Margolis SA, Park E. Stability of ascorbic acid in solutions stored in
autosampler vials. Clin Chem 2001;47:14634.
62. Drammeh BS, Schleicher RL, Pfeiffer CM, Jain RB, Zhang M, Nguyen
PH. Effects of delayed specimen processing and freezing on serum
concentrations of selected nutritional indicators. Clin Chem 2008;54:
188391.
63. Fraser CG, Hyltoft PP, Libeer JC, Ricos C. Proposals for setting generally applicable quality goals solely based on biology. Ann Clin Biochem 1997;34:812.

Downloaded from www.ajcn.org by on November 6, 2009

17. Seddon JM. Multivitamin-multimineral supplements and eye disease:

age-related macular degeneration and cataract. Am J Clin Nutr 2007;
85(suppl):304S7S.
18. Milton RC, Sperduto RD, Clemons TE, Ferris FL III. Centrum use and
progression of age-related cataract in the Age-Related Eye Disease
Study: a propensity score approach. AREDS report No. 21. Ophthalmology 2006;113:126470.
19. Yoshida M, Takashima Y, Inoue M, et al. Prospective study showing that
dietary vitamin C reduced the risk of age-related cataracts in a middle-aged
Japanese population. Eur J Nutr 2007;46:11824.
20. Hampl JS, Taylor CA, Johnston CS. Vitamin C deficiency and depletion
in the United States: the Third National Health and Nutrition Examination Survey, 1988 to 1994. Am J Public Health 2004;94:8705.
21. National Center for Health Statistics. Continuous NHANES web tutorial. 2008. Available from: http://www.cdc.gov/nchs/tutorials/Nhanes/
index_current.htm (cited 8 June 2009).
22. Ezzati TM, Massey JT, Waksberg J, Chu A, Maurer KR. Sample design:
Third National Health and Nutrition Examination Survey. Vital Health
Stat 2 1992;(113):135.
23. Anonymous. Plan and operation of the Third National Health and Nutrition Examination Survey, 1988-94. Series 1: programs and collection
procedures. Vital Health Stat 1 1994;1:1407.
24. McCoy LF, Bowen MB, Xu M, Chen H, Schleicher RL. Improved
HPLC assay for measuring serum vitamin C with 1-methyluric acid used
as an electrochemically active internal standard. Clin Chem 2005;51:
10624.
25. Gunter EW, Lewis BG, Koncikowski SM. Laboratory procedures used
for the Third National Health and Nutrition Examination Survey
(NHANES III), 19881994. 1996. Available from: http://www.cdc.
gov/nchs/data/nhanes/nhanes3/cdrom/nchs/manuals/labman.pdf (cited 8
June 2009).
26. Margolis SA, Duewer DL. Measurement of ascorbic acid in human
plasma and serum: stability, intralaboratory repeatability, and interlaboratory reproducibility. Clin Chem 1996;42:125762.
27. Pirkle JL, Flegal KM, Bernert JT, Brody DJ, Etzel RA, Maurer KR.
Exposure of the US population to environmental tobacco smoke. JAMA
1996;275:123340.
28. Bernert JT Jr, Turner WE, Pirkle JL, et al. Development and validation
of a sensitive method for determination of serum cotinine in smokers
and nonsmokers by liquid chromatography/atmospheric pressure ionization tandem mass spectrometry. Clin Chem 1997;43:228191.
29. Kuczmarski RJ, Flegal KM. Criteria for definition of overweight in
transition: background and recommendations for the United States. Am
J Clin Nutr 2000;72:107481.
30. National Center for Health Statistics. NHANES 20032004. Documentation, codebook, and frequencies. Demographic variables and
sample weights. 2005. Available from: http://www.cdc.gov/nchs/data/
nhanes/nhanes_03_04/demo_c.pdf (cited 8 June 2009).
31. Subcommittee on Interpretation and Use of Dietary Reference Intakes
and the Standing Committee on the Scientific Evaluation of Dietary
Reference Intakes. Dietary Reference Intakes: applications in dietary
planning. Washington, DC: National Academy Press, 2003.
32. Wolter KM. Introduction to variance estimation. New York, NY:
Springer-Verlag, 1985.
33. Korn EL, Graubard BI. Analysis of health surveys. New York, NY: John
Wiley and Sons Inc., 1999.
34. Agresti A. Categorical data analysis. Hoboken, NJ: John Wiley & Sons,
2002.
35. Klein RJ, Schoenborn CA. Age adjustment using the 2000 projected
U.S. population. Healthy People 2010 Stat Notes 2001;(20):110.
36. Miller RG. Developments in multiple comparisons 19661976. J Am
Stat Assn 1977;72:77988.
37. Winer BJ. Statistical principles in experimental design. New York, NY:
McGraw-Hill Companies, 1971:878 (Table C.10).
38. Food and Nutrition Board, Institute of Medicine. Vitamin C. In: DRI
Dietary Reference Intakes for vitamin C, vitamin E, selenium, and
carotenoids. Washington, DC: National Academy Press, 2000:95185.
39. National Center for Health Statistics. NHANES analytic guidelines. 2008.
Available from: http://www.cdc.gov/nchs/data/nhanes/nhanes_03_04/
nhanes_analytic_guidelines_dec_2005.pdf (cited 8 June 2009).
40. Padayatty SJ, Katz A, Wang Y, et al. Vitamin C as an antioxidant: evaluation of its role in disease prevention. J Am Coll Nutr 2003;22:1835.
41. Johnston CS. Vitamin C. In: Bowman BA, Russell RM, eds. Present
knowledge in nutrition. Washington, DC: ILSI, 2006:23341.

1263